This special issue is the ﬁrst of a three part series that focuses on emerging tech-
nologies for reliable engineering analysis and design. The papers represent selected
contributions to the National Science Foundation workshop on Reliable Engi-
neering Computing held September 2004 at the Center for Reliable Engineering
Computing, Georgia Tech Savannah.
One of the main objectives of engineering practice is to achieve reliable and
robust designs. However, the quality and reliability of the design tools plays a
main role in achieving this objective. Design tools include mathematical models,
algorithms, and numerical methods for analysis, optimization, and system simula-
Prediction of an engineering system behavior depends to a large extend on
the veriﬁcation and validation processes, where veriﬁcation is understood as the
process of evaluating the quality of the computed results in reﬂecting the behavior
of a mathematical model, and validation is the process of evaluating the reliability
of a mathematical model in reﬂecting reality.
In this context, reliable engineering computing, as we understand it, requires
that computing systems accommodate several sources of uncertainty and errors
with a focus on self-verifying methods. In the case of a mechanical system model,
uncertainties can originate from:
1) The appropriateness of the mathematical model to describe the physical system;
2) The discretization of the mathematical model into a computational framework;
3) The inexact knowledge of input parameters of a problem; and
4) Errors introduced by the nature of computer ﬁnite arithmetic.
A reliable engineering analysis must include all of the above in providing both
solutions and measures of the reliability of the results provided. To date, no all-
encompassing framework exists for reliable engineering computing. This ﬁeld will
only progress through interdisciplinary research that enables one to addresses the
different aspects of reliable engineering computing in both analysis and design.
Reliable Computing (2006)